CN114776304A - Method, device, equipment and medium for identifying abnormal deep sea mineral areas - Google Patents
Method, device, equipment and medium for identifying abnormal deep sea mineral areas Download PDFInfo
- Publication number
- CN114776304A CN114776304A CN202210605551.5A CN202210605551A CN114776304A CN 114776304 A CN114776304 A CN 114776304A CN 202210605551 A CN202210605551 A CN 202210605551A CN 114776304 A CN114776304 A CN 114776304A
- Authority
- CN
- China
- Prior art keywords
- target
- abnormal
- data
- area
- mineral
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 180
- 229910052500 inorganic mineral Inorganic materials 0.000 title claims abstract description 153
- 239000011707 mineral Substances 0.000 title claims abstract description 153
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000001514 detection method Methods 0.000 claims abstract description 13
- 229910052729 chemical element Inorganic materials 0.000 claims description 89
- 238000001228 spectrum Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000000513 principal component analysis Methods 0.000 claims description 6
- 238000007621 cluster analysis Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- 230000002547 anomalous effect Effects 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims description 2
- 238000009826 distribution Methods 0.000 description 26
- 229910052761 rare earth metal Inorganic materials 0.000 description 15
- 150000002910 rare earth metals Chemical class 0.000 description 12
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N Silicium dioxide Chemical compound O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 10
- GWEVSGVZZGPLCZ-UHFFFAOYSA-N Titan oxide Chemical compound O=[Ti]=O GWEVSGVZZGPLCZ-UHFFFAOYSA-N 0.000 description 10
- 230000003595 spectral effect Effects 0.000 description 7
- 229910052692 Dysprosium Inorganic materials 0.000 description 5
- 229910052693 Europium Inorganic materials 0.000 description 5
- 229910052688 Gadolinium Inorganic materials 0.000 description 5
- 229910052689 Holmium Inorganic materials 0.000 description 5
- 229910052779 Neodymium Inorganic materials 0.000 description 5
- 229910052772 Samarium Inorganic materials 0.000 description 5
- 229910052771 Terbium Inorganic materials 0.000 description 5
- 229910052775 Thulium Inorganic materials 0.000 description 5
- 229910052769 Ytterbium Inorganic materials 0.000 description 5
- PNEYBMLMFCGWSK-UHFFFAOYSA-N aluminium oxide Inorganic materials [O-2].[O-2].[O-2].[Al+3].[Al+3] PNEYBMLMFCGWSK-UHFFFAOYSA-N 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 5
- 229910052681 coesite Inorganic materials 0.000 description 5
- 229910052593 corundum Inorganic materials 0.000 description 5
- 229910052906 cristobalite Inorganic materials 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 229910052733 gallium Inorganic materials 0.000 description 5
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 5
- 229910052746 lanthanum Inorganic materials 0.000 description 5
- 229910052745 lead Inorganic materials 0.000 description 5
- VASIZKWUTCETSD-UHFFFAOYSA-N manganese(II) oxide Inorganic materials [Mn]=O VASIZKWUTCETSD-UHFFFAOYSA-N 0.000 description 5
- 229910052759 nickel Inorganic materials 0.000 description 5
- NOTVAPJNGZMVSD-UHFFFAOYSA-N potassium monoxide Inorganic materials [K]O[K] NOTVAPJNGZMVSD-UHFFFAOYSA-N 0.000 description 5
- 239000013049 sediment Substances 0.000 description 5
- 239000000377 silicon dioxide Substances 0.000 description 5
- 229910052682 stishovite Inorganic materials 0.000 description 5
- 229910052712 strontium Inorganic materials 0.000 description 5
- 229910052905 tridymite Inorganic materials 0.000 description 5
- 229910001845 yogo sapphire Inorganic materials 0.000 description 5
- 229910052727 yttrium Inorganic materials 0.000 description 5
- 229910052725 zinc Inorganic materials 0.000 description 5
- 229910052726 zirconium Inorganic materials 0.000 description 5
- KKCBUQHMOMHUOY-UHFFFAOYSA-N Na2O Inorganic materials [O-2].[Na+].[Na+] KKCBUQHMOMHUOY-UHFFFAOYSA-N 0.000 description 4
- 238000012896 Statistical algorithm Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 229910052706 scandium Inorganic materials 0.000 description 4
- ODINCKMPIJJUCX-UHFFFAOYSA-N calcium oxide Inorganic materials [Ca]=O ODINCKMPIJJUCX-UHFFFAOYSA-N 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 229910052802 copper Inorganic materials 0.000 description 3
- 238000013480 data collection Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- CPLXHLVBOLITMK-UHFFFAOYSA-N magnesium oxide Inorganic materials [Mg]=O CPLXHLVBOLITMK-UHFFFAOYSA-N 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 229910052684 Cerium Inorganic materials 0.000 description 2
- 241000801593 Pida Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 229910052804 chromium Inorganic materials 0.000 description 2
- 238000007417 hierarchical cluster analysis Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012847 principal component analysis method Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 229910052720 vanadium Inorganic materials 0.000 description 2
- 241001506304 Kadsura japonica Species 0.000 description 1
- 241000145637 Lepturus Species 0.000 description 1
- 235000005775 Setaria Nutrition 0.000 description 1
- 241000232088 Setaria <nematode> Species 0.000 description 1
- 230000033558 biomineral tissue development Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21C—MINING OR QUARRYING
- E21C50/00—Obtaining minerals from underwater, not otherwise provided for
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mining & Mineral Resources (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geochemistry & Mineralogy (AREA)
- Geology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the application provides a method, a device, equipment and a medium for identifying a deep sea mineral abnormal area, wherein the method comprises the following steps: acquiring a pre-classified image and geochemical data corresponding to the pre-classified image, wherein the pre-classified image is obtained by preliminarily identifying an abnormal area, and the geochemical data is obtained by detecting deep sea mineral products of a target sea area; and inputting the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification to obtain a target abnormal region classification result so as to finish further detection of the target sea area. Through some embodiments of the method and the device, the geochemical data characteristics of the target sea area minerals can be comprehensively considered, and then the more accurate abnormal deep sea mineral area is identified through the target abnormal area identification model, so that the detection efficiency is improved.
Description
Technical Field
The embodiment of the application relates to the field of deep sea detection, in particular to a method, a device, equipment and a medium for identifying a deep sea mineral abnormal area.
Background
With the new discovery of fewer and fewer superficial ores on land, mineral exploration has become a trend towards deep sea. The deep sea mineral products have the characteristics of large buried depth, weak mineralization information and the like, and the exploration difficulty of the deep sea mineral products is increased due to the unpredictability of offshore operation.
In the related technology, the traditional statistical algorithm is used for calculating the geochemical data of the deep sea minerals to obtain the abnormal area of the target sea area. However, the delineated anomalous regions have a large uncertainty since conventional statistical algorithms cannot take into account the intrinsic characteristics of the mineral geochemical data.
Therefore, how to accurately identify the abnormal deep sea mineral areas becomes a problem to be solved.
Disclosure of Invention
The embodiment of the application provides a method, a device, equipment and a medium for identifying a deep sea mineral abnormal area, and at least geochemical data characteristics of a target sea area mineral can be comprehensively considered through some embodiments of the application, so that the more accurate deep sea mineral abnormal area is identified through a target abnormal area identification model, and the detection efficiency is improved.
In a first aspect, the present application provides a method for identifying an abnormal region of deep sea mineral production, the method comprising: acquiring a pre-classified image and geochemical data corresponding to the pre-classified image, wherein the pre-classified image is obtained by preliminarily identifying an abnormal area, and the geochemical data is obtained by detecting deep sea mineral products in a target sea area; and inputting the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification to obtain a target abnormal region classification result so as to finish further detection of the target sea area.
Therefore, compared with the method for directly using the multi-fractal delineation of the abnormal region in the related technology, the method for delineating the abnormal region in the target sea area has the advantages that the pre-classified image is input into the identification model of the abnormal region in the target for further identification, the more accurate abnormal region can be delineated, and the efficiency of detecting the target sea area can be further improved.
With reference to the first aspect, in one embodiment of the present application, the pre-classified image includes a normal region and the abnormal region, the content values of each chemical element in the geochemical data of the normal region are all within a set threshold range, and the content values of at least some chemical elements in the abnormal region are not within the set threshold range; before the inputting the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification and obtaining a target abnormal region classification result, the method further comprises the following steps: obtaining a pre-classified sample image; inputting at least a pre-classified sample image and geochemical data corresponding to the pre-classified sample image into an abnormal area recognition model to be trained for training to obtain the target abnormal area recognition model; wherein the abnormal region marked by the label in the pre-classified sample image belongs to a sub-region of the abnormal region in the pre-classified image.
Therefore, by training the abnormal area identification model to be trained, the target abnormal area identification model with more precisely identified abnormal areas can be obtained, and further detection of deep-sea mineral products in the target area can be realized.
With reference to the first aspect, in an embodiment of the present application, before the acquiring the pre-classified sample image, the method further includes: dividing the pre-classified image into M regions, wherein M is an integer greater than 1; confirming that the M areas comprise N abnormal areas, wherein N is an integer greater than or equal to 1, and N is less than or equal to M; and marking each abnormal region in the N abnormal regions as an enriched region or a depleted region according to the content of any chemical element included in each abnormal region, and obtaining the pre-classified sample image, wherein the content of the chemical element in the enriched region is greater than a first threshold value, the content of the chemical element in the depleted region is less than a second threshold value, and the second threshold value is less than the first threshold value.
Therefore, by dividing the abnormal regions in the pre-classified images into more precise abnormal regions to obtain the pre-classified sample images, the embodiment of the application can enable the abnormal region identification model to be trained to learn more precise abnormal region characteristics.
With reference to the first aspect, in one embodiment of the present application, before the acquiring the pre-classified image and the geochemical data corresponding to the pre-classified image, the method further comprises: acquiring original mineral data, wherein the original mineral data is geochemical data corresponding to deep sea minerals of the target sea area; dividing the original mineral data into a plurality of groups according to the attribute of each chemical element in the original mineral data to obtain a target classification result, wherein one group comprises at least one chemical element; and according to the target classification result, preliminarily identifying the normal area and the abnormal area corresponding to the original mineral data to obtain the pre-classified image.
Therefore, the normal area and the abnormal area are preliminarily identified through the attributes and the correlation among the chemical elements, the pre-classification image is obtained, data with high value can be provided for subsequent re-identification through the neural network, and therefore the accurate target abnormal area classification result can be obtained.
With reference to the first aspect, in an embodiment of the present application, the dividing the raw mineral data into multiple groups according to attributes of chemical elements in the raw mineral data to obtain a target classification result includes: dividing the original mineral data into a plurality of groups through a multi-fractal algorithm to obtain a first classification result; performing principal component analysis on the original mineral data, and dividing the original mineral data into a plurality of groups to obtain a second classification result; performing cluster analysis on the original mineral data, dividing the original mineral data into a plurality of groups, and obtaining a third classification result; and confirming that the first classification result, the second classification result and the third classification result are the same to obtain the target classification result.
Therefore, the original mineral data are grouped by multiple methods, and multiple results can be verified mutually, so that the classification accuracy is improved.
With reference to the first aspect, in an embodiment of the present application, the dividing the original mineral data into multiple groups through a multi-fractal algorithm to obtain a first classification result includes: calculating a multi-fractal spectrum curve corresponding to each chemical element; and dividing the original mineral data into a plurality of groups according to the multi-fractal spectrum curve to obtain the first classification result.
Therefore, the multi-fractal spectral curve corresponding to each chemical element is calculated, the intrinsic space structure of each chemical element can be effectively reflected, and the classification accuracy can be improved.
In a second aspect, the present application provides a system for deep sea mineral anomaly identification, the system comprising: the geochemical data acquisition equipment is used for detecting the deep-sea mineral products of the target sea area to obtain the original mineral product data of the target sea area; a server side configured to execute the method according to any embodiment of the first aspect according to the original mineral data to obtain a target abnormal area classification result; and the deep sea exploration equipment is used for further detecting the target sea area according to the target abnormal area division result.
In a third aspect, the present application provides an apparatus for deep sea mineral anomaly identification, the apparatus comprising: the acquisition module is configured to acquire a pre-classified image and geochemical data corresponding to the pre-classified image, wherein the pre-classified image is obtained by preliminarily identifying an abnormal area, and the geochemical data is obtained by detecting deep sea mineral products of a target sea area; and the identification module is configured to input the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification, and obtain a target abnormal region division result so as to finish further detection of the target sea area.
With reference to the third aspect, in one embodiment of the present application, the pre-classification image includes a normal region and the abnormal region, the content value of each chemical element in the geochemical data of the normal region is within a set threshold range, and the content value of at least part of the chemical elements in the abnormal region is not within the set threshold range; the identification module is configured to: obtaining a pre-classified sample image; inputting at least a pre-classified sample image and geochemical data corresponding to the pre-classified sample image into an abnormal area identification model to be trained for training to obtain the target abnormal area identification model; wherein the abnormal region marked by the label in the pre-classified sample image belongs to a sub-region of the abnormal region in the pre-classified image.
With reference to the third aspect, in one embodiment of the present application, the identification module is configured to: dividing the pre-classified image into M areas, wherein M is an integer greater than 1; confirming that the M areas comprise N abnormal areas, wherein N is an integer greater than or equal to 1, and N is less than or equal to M; and marking each abnormal region in the N abnormal regions as an enriched region or a depleted region according to the content of any chemical element included in each abnormal region, and obtaining the pre-classified sample image, wherein the content of the chemical element in the enriched region is greater than a first threshold value, the content of the chemical element in the depleted region is less than a second threshold value, and the second threshold value is less than the first threshold value.
With reference to the third aspect, in an embodiment of the present application, the obtaining module is configured to: acquiring original mineral data, wherein the original mineral data is geochemical data corresponding to deep sea minerals of the target sea area; dividing the original mineral data into a plurality of groups according to the attribute of each chemical element in the original mineral data to obtain a target classification result, wherein one group comprises at least one chemical element; and according to the target classification result, preliminarily identifying the normal area and the abnormal area corresponding to the original mineral data to obtain the pre-classified image.
With reference to the third aspect, in one embodiment of the present application, the identification module is configured to: dividing the original mineral data into a plurality of groups through a multi-fractal algorithm to obtain a first classification result; performing principal component analysis on the original mineral data, and dividing the original mineral data into a plurality of groups to obtain a second classification result; performing cluster analysis on the original mineral data, dividing the original mineral data into a plurality of groups, and obtaining a third classification result; and confirming that the first classification result, the second classification result and the third classification result are the same to obtain the target classification result.
With reference to the third aspect, in one embodiment of the present application, the identification module is configured to: calculating a multi-fractal spectrum curve corresponding to each chemical element; and dividing the original mineral data into a plurality of groups according to the fractal spectrum curve to obtain the first classification result.
In a fourth aspect, the present application provides an electronic device, comprising: a processor, a memory, and a bus; the processor is connected to the memory via the bus, and the memory stores computer readable instructions for implementing the method according to any of the embodiments of the first aspect when the computer readable instructions are executed by the processor.
In a fifth aspect, the present application provides a computer readable storage medium having stored thereon a computer program which, when executed, implements a method as in any of the embodiments of the first aspect.
Drawings
FIG. 1 is a schematic diagram of a system for identifying an abnormal area of deep sea mineral deposits according to an embodiment of the present application;
FIG. 2 is a flow chart of a method for identifying an abnormal deep sea mineral area according to an embodiment of the application;
FIG. 3 is a diagram illustrating a pre-classified image according to an embodiment of the present application;
FIG. 4 is a second flowchart of a method for identifying an abnormal area of deep sea minerals according to an embodiment of the present application;
FIG. 5 is a third flowchart of a method for identifying an abnormal deep sea mineral area according to an embodiment of the present application;
FIG. 6 is a block diagram of the device for identifying the abnormal deep sea mineral area according to the embodiment of the application;
fig. 7 is a schematic diagram illustrating a composition of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. The components of the embodiments of the present application, as generally described and illustrated in the figures herein, could be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, as presented in the figures, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.
In some embodiments of the present application, the pre-classified image is re-identified by the target abnormal area identification model to obtain a target abnormal area division result, wherein the target abnormal area division result defines an abnormal area of the target sea area, so that the deep sea detection device further detects the defined abnormal area mineral. For example: in some embodiments of the application, firstly, the geochemical data of the target sea area is preliminarily identified to obtain a pre-classified image, then the pre-classified image and the geochemical data are input into a target abnormal area identification model to be identified again, and a target abnormal area classification result is obtained. Through some embodiments of the method and the device, the geochemical data characteristics of the target sea area minerals can be comprehensively considered, and then the more accurate abnormal deep sea mineral area is identified through the target abnormal area identification model, so that the detection efficiency is improved.
The method steps in the embodiments of the present application are described in detail below with reference to the accompanying drawings.
FIG. 1 provides a schematic diagram of a system for deep sea mineral anomaly identification including a geochemical data collection facility 110, a service 120 and a deep sea exploration facility 130 in some embodiments of the present application. Specifically, the geochemical data collection device 110 performs geological exploration in the deep sea of the target sea area and obtains the raw mineral data (i.e., geochemical data) of the target sea area after collecting a sample. Then, the original mineral data is input into the server 120, and the server 120 executes the method for identifying the deep sea mineral abnormal area to obtain the target abnormal area classification result. Finally, the deep sea exploration equipment 130 performs deep sea exploration based on the target abnormal region demarcated by the result of the target abnormal region, and it is understood that deep sea exploration includes mineral exploration, soil exploration, biological exploration, and the like.
Different from the embodiment of the application, in the related technology, the geochemical data of the deep sea minerals are calculated by using the traditional statistical algorithm to obtain the abnormal area of the target sea area. However, the delineated anomalous regions have a large uncertainty since conventional statistical algorithms cannot take into account the intrinsic characteristics of the mineral geochemical data. The abnormal region is defined by re-identifying the pre-classified image through the target abnormal region identification model, so that the embodiment of the application can extract the internal characteristics of the pre-classified image and the corresponding geochemical data, and further obtain an accurate target abnormal region classification result.
The following will explain the implementation steps of a method for identifying a deep sea mineral abnormal area in the embodiment of the application.
At least to solve the problems in the background art, as shown in fig. 2, some embodiments of the present application provide a method for identifying an abnormal deep-sea mineral area, the method including:
s210, acquiring a pre-classified image and geochemical data corresponding to the pre-classified image.
It is understood that the geochemical data is obtained by exploring the deep sea mineral deposits in the target sea area. For example, geochemical data for deep sea sediments in the sea basin of kaja fischeri include the following chemical elements: ba. Sc, Y, La, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Pr, Er, Tm, Yb, Lu, REY, P2O5、Na2O、Co、Ce、Cr、Pb、Ga、SiO2、MnO、TiO2、Al2O3、Fe2O3、Cu、V、MgO、CaO、K2O, Ni, Zn, Sr and Zr.
It should be noted that the pre-classified image in S210 includes a normal area and an abnormal area. Specifically, the content of each chemical element in the geochemical data of the normal zone is within a set threshold range, and the content of part of the chemical elements or the content of all the chemical elements in the abnormal zone is not within the set threshold range. For example, if the target marine deep sea mineral is set to have a chemical element Ba content of 30% to 50%, the element content of Ba element in the normal region in the pre-classification image is 45%, and the element content of Ba element in the abnormal region is 80%, that is, the element content of Ba element in the abnormal region is not within the set threshold.
It should be noted that the normal zone is called a background zone in the deep sea geological exploration field.
In one embodiment of the present application, before S210, a pre-classification image is obtained from the original mineral data of the target sea area, which includes the following specific steps:
s1: raw mineral data is obtained.
For example, exploration is performed for deep-sea minerals in the target sea area to obtain geochemical data (i.e., raw mineral data) of the target sea area. The server executes S1 to obtain the original mineral data.
It is understood that the above method for obtaining raw mineral data is only an example, and the application does not limit the method for obtaining raw mineral data.
S2: according to the attribute of each chemical element in the original mineral data, dividing the original mineral data into a plurality of groups to obtain a target classification result, wherein one group comprises at least one chemical element.
S201: original mineral data are divided into a plurality of groups through a multi-fractal algorithm, and a first classification result is obtained.
That is, the first method for classifying raw mineral data is a multi-fractal algorithm, i.e., a multi-fractal spectral curve corresponding to each chemical element is calculated, and then the raw mineral data is classified according to each parameter characterized in the multi-fractal spectral curve.
It will be appreciated that, spatially, a multifractal is an entanglement of single fractals, which describes a metric, typically in a two-dimensional or three-dimensional representation, referred to as a multifractal if the metric has spatial or statistical self-similarity.
The method comprises the following specific steps:
first, a fractal spectrum curve corresponding to each chemical element is calculated.
That is, a moment method is used to calculate a fractal spectrum curve corresponding to each chemical element.
Specifically, when constructing the fractal spectrum curve, q is an arbitrary number, and the value of q is [ -10,10 [ -10]Interval is 1 when epsilonsAnd smaller, appears as a straight line. Calculating the slope tau (q) of the straight line by adopting least square method fitting, and then establishing a singularity index and qAnd finally obtaining a multi-fractal spectrum f (alpha) through Lagrange transformation to obtain a multi-fractal spectrum curve.
And then, dividing the original mineral data into a plurality of groups according to the multi-fractal spectrum curve to obtain a first classification result.
Specifically, in the embodiment of the present application, an asymmetric index R is used to measure the deviation degree between the fractal spectrum curve of each chemical element and the symmetric fractal spectrum curve, where the value range of R is [ -1,1], and the calculation method of the R value is shown in the following formula:
ΔαL＝|αmin-α0|
ΔαR＝|αmax-α0|
R＝(ΔαL-AαR)/(ΔαL+ΔαR)
wherein R represents an asymmetry index, Δ αLThe result of expressing the difference between the minimum singular index and the singular index of 2 and then taking the absolute value, Δ αRA result of expressing the difference between the maximum singularity index and the singularity index of 2 and then taking the absolute value, alphaminRepresenting the minimum value of the singularity index, αmaxRepresenting the maximum value of the singularity index, α0Represents a value of α when q is 0. It can be understood that when R is 0, it means that the fractal spectrum curve of the chemical element is completely symmetrical on both sides, when R is greater than 0, it means that the fractal spectrum curve of the chemical element is biased to the left, and when R is less than 0, it means that the fractal spectrum curve of the chemical element is biased to the right.
And then, analyzing by combining the multi-fractal spectrum curve and the R value of each element, and dividing the original mineral data into a plurality of groups. Specifically, the chemical elements included in the first group are: ba. Sc, Y, La, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Pr, Er, Tm, Yb, Lu, REY, P2O5And Na2And (O). It can be understood that the first group has a relatively large Δ α and a corresponding large Δ f (α), and shows a strong multi-fractal characteristic, the R values of the elements in the group are all larger than 0, and the multi-fractal spectrum curve is biased to the left. The chemical elements included in the second group are: co, Ce, Cr, Pb, Ga, SiO2、MnO、TiO2、Al2O3、Fe2O3Cu and V. It is understood that the R values of the chemical elements in the second group are less than 0, the fractal spectrum curves are biased to the right, and the set of indices have narrower continuous fractal spectrum curves, showing very weak fractal characteristics. The third group comprises the following chemical elements: MgO, CaO, K2O, Ni, Zn, Sr and Zr. It is understood that the absolute value of the R value corresponding to each chemical element in the third group is less than 0.5, and the Δ α in the group is relatively small, i.e. it means that each chemical element in the group is a simple single fractal.
It can be understood that the multifractal spectrum curve can effectively reflect the intrinsic spatial structure of each chemical element, and therefore, three relatively constant parameters, namely, Δ α, R value and coefficient of variation, can be used to characterize the intrinsic spatial characteristics of each chemical element.
Therefore, the multi-fractal spectral curve corresponding to each chemical element is calculated, the intrinsic space structure of each chemical element can be effectively reflected, and the classification accuracy can be improved.
S202: and performing principal component analysis on the original mineral data, and dividing the original mineral data into a plurality of groups to obtain a second classification result.
That is, the second way to classify raw mineral data is principal component analysis. It can be understood that the principal component analysis method is to try to recombine a plurality of original indexes with certain correlation into a group of new unrelated comprehensive indexes to replace the original indexes.
Specifically, the principal component analysis method comprises the following main steps: standardizing the content of each chemical element, judging the correlation among the chemical elements, determining the number m of the principal components, determining the expression of the principal component Fi, naming the principal component Fi, and finally dividing the original mineral data into a plurality of groups to obtain a second classification result. In the second classification result, the chemical elements included in the first main component are: ba. Sc, Y, La, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Pr, Er, Tm, Yb, Lu, REY, P2O5And Na2O, the second main component comprisesThe chemical elements of (A) are: co, Ce, Cr, Pb, Ga, SiO2、MnO、TiO2、Al2O3、Fe2O3Cu and V, and the third main component comprises the following chemical elements: MgO, CaO, K2O, Ni, Zn, Sr and Zr.
S203: and performing cluster analysis on the original mineral data, dividing the original mineral data into a plurality of groups, and obtaining a third classification result.
As an embodiment of the application S203, feature parameter analysis needs to be performed on each chemical element in the original mineral data, that is, skewness, kurtosis, standard deviation, and variation coefficient of each chemical element are calculated, then, hierarchical clustering analysis is performed according to the plurality of feature parameters, the original mineral data is divided into a plurality of groups, and a third classification result is obtained.
It is understood that skewness is a measure of the direction and degree of skew of the statistical data distribution, and is a numerical characteristic of the degree of asymmetry of the statistical data distribution, wherein skewness definition includes normal distribution (skewness is 0), left-skewed distribution (also called negative-skewed distribution, whose skewness is less than 0), and right-skewed distribution (also called positive-skewed distribution, whose skewness is greater than 0).
The kurtosis is also called a kurtosis coefficient, and characterizes the characteristic number of the probability density distribution curve of the peak value height at the average value. Intuitively, the kurtosis reflects the sharpness of the peak. The kurtosis calculation method of the random variable comprises the following steps: the ratio of the fourth-order central moment of the random variable to the square of the variance. Kurtosis includes normal distribution (kurtosis value of 3), thick tail (kurtosis value >3), and thin tail (kurtosis value < 3). Kurtosis is similar to skewness and is a statistic for describing the steepness of all value distribution forms in the population. This statistic needs to be compared with the normal distribution, with a kurtosis of 0 indicating that the overall data distribution is as steep as the normal distribution; the kurtosis is larger than 0, which means that the overall data distribution is steeper than the normal distribution and is a sharp peak; a kurtosis of less than 0 indicates that the overall data distribution is relatively flat compared to a normal distribution, which is a flat-topped peak. The larger the absolute value of kurtosis is, the larger the degree of difference between the steepness of the distribution form and the normal distribution is.
The coefficient of variation is an important index for evaluating the degree of element differentiation, and can be obtained by dividing the standard deviation by the average value.
As a specific embodiment of the present application, taking a plurality of chemical elements in the geochemical data of the target sea area as an example, the result of calculating the characteristic parameters is exemplarily shown, as shown in table 1:
TABLE 1 calculation of characteristic parameters of chemical elements
After calculating to obtain the characteristic parameter calculation results of the chemical elements, performing hierarchical clustering analysis on the chemical elements according to the characteristic parameters of the chemical elements, and then dividing all the chemical elements in the original mineral data into a plurality of groups, wherein the chemical elements in the first group are as follows: ba. Sc, Y, La, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Pr, Er, Tm, Yb, Lu, REY, P2O5And Na2O, the chemical elements included in the second group are: co, Ce, Cr, Pb, Ga, SiO2、MnO、TiO2、Al2O3、Fe2O3Cu and V, the third group comprising the chemical elements: MgO, CaO, K2O, Ni, Zn, Sr and Zr.
As another embodiment of the present application S203, R-type clustering is used to group chemical elements to obtain a third classification result.
It is understood that R-type clustering is one of the clustering methods. And classifying according to the degree of correlation between different variables. If the variables are more and the correlation is stronger, the R-type clustering method can be used for clustering the variables into a plurality of large classes, the same variables have stronger correlation, the correlation degree between different classes of variables is low, a typical variable can be found out from the similar variables to be used as a representative, and finally the number of the variables is reduced to achieve the purpose of reducing the dimension.
It is to be understood that the present application does not limit the execution sequence of the above-mentioned S201 to S203. That is, as a specific embodiment of the present application, S202 may be executed first, and then S201 and S203 may be executed, and as another specific embodiment of the present application, S201, S202, and S203 may be executed simultaneously.
S204: and confirming that the first classification result, the second classification result and the third classification result are the same to obtain a target classification result.
That is, after the original mineral data is grouped by the above methods, a plurality of classification results, that is, the first classification result, the second classification result, and the third classification result, can be obtained. Then judging whether the three classification results are the same, and if so, taking any one classification result as a target classification result; if one or more of the three classification results are different, recalculation is needed until the results are unified.
Therefore, the original mineral data are grouped by multiple methods, and multiple results can be verified mutually, so that the classification accuracy is improved.
S3: and according to the target classification result, preliminarily identifying a normal area and an abnormal area corresponding to the original mineral data to obtain a pre-classification image.
That is, the target classification result obtained in S2 is used, the multi-fractal filtering technique and the singularity index model are used to analyze the geochemical anomaly of the target sea area, the anomaly information which is weakly difficult to identify is extracted, and the spatial distribution area of the anomaly region is defined. It can be understood that the advantage of the multi-fractal filtering and the singularity index model is that the multi-fractal filtering and the singularity index model can extract abnormal weak information and finely calculate the abnormal distribution range.
Specifically, an S-A method is applied to perform multi-fractal processing on A target classification result, namely S-A log fitting is performed on chemical elements of each group in the target classification result, and A demarcation point is determined based on A least square method principle. Generally, the S-A log graph is divided into 3 fitting straight lines or 4 fitting straight lines (for example, the fitting straight lines corresponding to the chemical elements included in the first group are represented by 3 or 4 straight lines, where 1 straight line corresponds to one fitting equation), and the embodiment of the present application finds that the precision is higher when 4 fitting straight lines are used after comparing 3 fitting straight lines and 4 fitting straight lines. Therefore, the target sea area in the application adopts a 4-segment method to separate the abnormal area from the normal area.
In a specific embodiment of the present application, taking the first group as an example, after performing log-log fitting on the first group, the energy spectral density of the first group after the first group is divided into 4 segments is 5-28, 28-646, 646-. In order to check the significance of the regression equation in each section, error check is carried out on each section, and fitting errors of each section are respectively 0, 0.001 and 0.002 through calculation. Two filters are defined with a threshold of 646: the outlier filter S <646 and the background filter S > 646, i.e., chemical elements with spectral densities above 646 are classified as background regions (i.e., normal regions), and chemical elements with spectral densities below 646 are classified as outlier regions.
Therefore, the normal area and the abnormal area are preliminarily identified through the attributes and the correlation among the chemical elements, the pre-classification image is obtained, data with high value can be provided for subsequent re-identification through the neural network, and therefore accurate target abnormal area distinguishing results can be obtained.
S220, inputting the geochemical data of the pre-classified images into a target abnormal area recognition model for re-recognition to obtain a target abnormal area classification result.
S1: the pre-classified image is divided into M regions, wherein M is an integer greater than 1.
That is, in order to enable the target abnormal region identification model to more effectively learn the features in the pre-classified images, the present application divides the pre-classified images. As a specific embodiment in the present application S1, as shown in fig. 3, a pre-classified image is divided into 4 regions.
S2: and confirming that the M areas comprise N abnormal areas, wherein N is an integer greater than or equal to 1, and N is less than or equal to M.
That is, as shown in fig. 3, of the above-described divided 4 areas, the abnormal areas, i.e., the first abnormal area and the second abnormal area, are found.
S3: and marking each abnormal area in the N abnormal areas as an enriched area or a depleted area respectively according to the content of any chemical element included in each abnormal area, and obtaining a pre-classification sample image.
It should be noted that the content of the chemical element in the rich region is greater than the first threshold, the content of the chemical element in the lean region is less than the second threshold, and the second threshold is less than the first threshold.
In other words, in the process of marking the sample, the content of any chemical element included in each abnormal region is acquired, the abnormal region is marked as an enriched region when the content of any chemical element is greater than a first threshold, and the abnormal region is marked as a depleted region when the content of any chemical element is less than a second threshold, so that a pre-classified sample image is acquired.
For example, as shown in fig. 3, if the content of Ba element in the first abnormal region is greater than the first threshold, the first abnormal region is marked as a rich region, and if the content of Fe element in the second abnormal region is less than the second threshold, the second abnormal region is marked as a lean region.
Therefore, the embodiment of the application divides the abnormal regions in the pre-classified images into more precise abnormal regions to obtain the pre-classified sample images, so that the abnormal region identification model to be trained can learn more precise abnormal region characteristics.
S4: and acquiring a pre-classified sample image.
S5: and inputting at least the pre-classified sample image and the geochemical data corresponding to the pre-classified sample image into the abnormal area recognition model to be trained for training to obtain the target abnormal area recognition model.
That is, the pre-classified sample image, the geochemical data corresponding to the pre-classified sample image, and the like are input into the abnormal region identification model to be trained for training, so that the abnormal region identification model to be trained can learn the characteristics of the marked finer abnormal regions in the pre-classified sample image.
In one embodiment of the present application, as shown in fig. 4, the pre-classified sample image and the geochemical data are input into the abnormal region identification model 410 to be trained for training, and the target abnormal region identification model 420 and the model parameters corresponding to the target abnormal region identification model are obtained. Then, the pre-classified images and the geochemical data corresponding to the pre-classified images are input into a target abnormal region identification model 420, the characteristics of the abnormal regions are extracted, more fine abnormal regions are defined, and finally, the target abnormal region classification result is obtained.
It will be appreciated that the labeled anomaly region in the pre-sorted sample image belongs to a sub-region of the anomaly region in the pre-sorted sample image, e.g., as shown in fig. 3, labeled anomaly region 301 belongs to a sub-region of the first anomaly region; the labeled anomaly region 302 belongs to a sub-region of the second anomaly region.
It should be noted that, in the present application, the target abnormal region identification model or the abnormal region identification model to be trained is established based on a convolutional neural network.
Therefore, by training the abnormal area identification model to be trained, the target abnormal area identification model with more precisely identified abnormal areas can be obtained, and further detection of deep-sea mineral products in the target area can be realized.
Therefore, the method starts from an innovative theoretical method, develops the traditional nonlinear geochemical information identification technology to a fractal transformation domain by combining the scale invariance of a fractal theory and a machine learning model of a deep learning hidden layer, and realizes the measurement of the anisotropy and the singularity of deep sea mineral information; on the other hand, starting from the actual problem of deep sea mineral products, the problems of site selection, weak and slow abnormity identification and extraction and the like in prediction and evaluation of mineral products at the deep part of the Kajiajiajiata basin (namely a target sea area) of the Western Pacific ocean are solved. Therefore, the rare earth mineral product mining information identification theory and method can be further enriched and developed, the multiscale coupling characteristics of the western Pacific ocean Pija javanica basin mining system can be deeply recognized, technical support and scientific basis are hopefully provided for deep sea mining of the western Pacific ocean, and the direction is pointed for ocean drilling.
In conclusion, according to the method, a western pacific skin jia he sea basin is taken as a target sea area, deep sea sediment geochemical data are used, the advantage that the weak information of the submarine mineral products can be identified through deep learning is utilized for identification, a multi-fractal model and a convolutional neural network model are adopted for carrying out feature identification on the deep sea rare earth to extract the abnormal fine morphology of elements, the distribution range of the rare earth resources is further defined, and a basis is provided for submarine mineral resource evaluation. The method can provide a new idea for geochemical exploration of the deep sea rare earth-rich sediments.
The foregoing describes a method for identifying a deep sea mineral abnormal region in an embodiment of the present application, and the following describes a specific embodiment of the deep sea mineral abnormal region identification in the present application.
In the related technology, the natural conditions of a target sea area are severe, the sea conditions are complex and changeable, the research degree is relatively low, the research on the aspect of deep sea rare earth mineral product identification is less, the spatial distribution range of the rare earth mineral product is unbalanced, the rare earth mineral product enrichment rules under different geological conditions are different, and therefore a unified rare earth mineral product identification index is difficult to establish, the reliability of defining an abnormal result by utilizing geochemical data is directly influenced, and the accuracy degree of later-stage mineral resource evaluation is influenced.
As one of various scenes, the western pacific pida sea basin is taken as a research area, geochemical data of deep sea sediment samples obtained through test analysis are taken as the basis, the rare earth resource enrichment rule is further explored on the basis of fully learning and absorbing research results of predecessors, and a multi-fractal model is adopted to carry out geochemical data (Ba, Sc, Y, La, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Pr, Er, Tm, Yb, Lu, REY and P) on various elements of the deep sea sediment of the pacific pida sea basin2O5、Na2O、Co、Ce、Cr、Pb、Ga、SiO2、MnO、TiO2、Al2O3、Fe2O3、Cu、V、MgO、CaO、K2O, Ni, Zn, Sr, and Zr) as convolution, and the obtained geochemical abnormal shapeThe neural network model inputs an information source to perform deep learning feature extraction, and the spatial distribution range of rare earth element enrichment is defined to improve the precision of rare earth resource evaluation.
As an embodiment of the present application, as shown in fig. 5, data collection and sorting are first performed S510 to obtain the geology, geophysical data 520, and geochemical data 540 of the kajia kadsura. Then, the abnormal area is preliminarily identified according to the geology of the Picjia Setaria basin, the geophysical data 520 and the geochemical data 540, and a pre-classification image 530 is obtained. Then, the pre-classified image 530, the pijiafischer sea basin geology, the geophysical data 520, and the geochemical data 540 are input into the target abnormal region identification model 420, and a target abnormal region classification result 550 is obtained. Then, performing S560 to verify with the known mineral distribution range, if not, continuing to perform the abnormal region identification by using the target abnormal region identification model 420, and if yes, performing S570 to identify the unknown rare earth-enriched region by circling.
Therefore, the embodiment of the application utilizes the known geochemical data to obtain the unknown element information, and improves the resolution and the recognition rate of the rare earth elements. Taking a pre-classified image obtained by a multi-fractal model as an input information source, performing feature extraction on deep sea rare earth elements by using a convolutional neural network model, identifying and extracting element abnormal information, and further identifying and delineating a rare earth mineral spatial distribution range.
The above describes a specific embodiment of the present application of the identification of the abnormal region of deep sea minerals, and the following describes a device of the present application of the identification of the abnormal region of deep sea minerals.
As shown in fig. 6, an apparatus 600 for identifying an abnormal area of deep sea minerals comprises: an acquisition module 610 and an identification module 620.
An obtaining module 610 configured to obtain a pre-classified image obtained by performing preliminary identification on an abnormal area and geochemical data corresponding to the pre-classified image obtained by detecting deep-sea mineral products of a target sea area.
And the recognition module 620 is configured to input the pre-classified image and the geochemical data into a target abnormal region recognition model for re-recognition, and obtain a target abnormal region classification result so as to complete further detection of the target sea area.
In one embodiment of the present application, the pre-classified image includes a normal region and the abnormal region, the content values of each chemical element in the geochemical data of the normal region are all within a set threshold range, and the content values of at least part of the chemical elements in the abnormal region are not within the set threshold range; the identification module 620 is configured to: obtaining a pre-classified sample image; inputting at least a pre-classified sample image and geochemical data corresponding to the pre-classified sample image into an abnormal area recognition model to be trained for training to obtain the target abnormal area recognition model; and the abnormal area marked by the label in the pre-classified sample image belongs to a sub-area of the abnormal area in the pre-classified image.
In one embodiment of the present application, the identification module 620 is configured to: dividing the pre-classified image into M regions, wherein M is an integer greater than 1; confirming that the M areas comprise N abnormal areas, wherein N is an integer greater than or equal to 1, and N is less than or equal to M; and marking each abnormal region in the N abnormal regions as an enriched region or a depleted region according to the content of any chemical element included in each abnormal region, and obtaining the pre-classified sample image, wherein the content of the chemical element in the enriched region is greater than a first threshold value, the content of the chemical element in the depleted region is less than a second threshold value, and the second threshold value is less than the first threshold value.
In one embodiment of the present application, the obtaining module 610 is configured to: acquiring original mineral data, wherein the original mineral data is geochemical data corresponding to deep sea minerals of the target sea area; dividing the original mineral data into a plurality of groups according to the attribute of each chemical element in the original mineral data to obtain a target classification result, wherein one group comprises at least one chemical element; and according to the target classification result, preliminarily identifying the normal area and the abnormal area corresponding to the original mineral data to obtain the pre-classified image.
In one embodiment of the present application, the identification module 620 is configured to: dividing the original mineral data into a plurality of groups through a multi-fractal algorithm to obtain a first classification result; performing principal component analysis on the original mineral data, and dividing the original mineral data into a plurality of groups to obtain a second classification result; performing cluster analysis on the original mineral data, dividing the original mineral data into a plurality of groups, and obtaining a third classification result; and confirming that the first classification result, the second classification result and the third classification result are the same to obtain the target classification result.
In one embodiment of the present application, the identification module 620 is configured to: calculating a multi-fractal spectrum curve corresponding to each chemical element; and dividing the original mineral data into a plurality of groups according to the multi-fractal spectrum curve to obtain the first classification result.
In this embodiment of the present application, the module shown in fig. 6 can implement each process in the method embodiments of fig. 1 to 5. The operations and/or functions of the respective modules in fig. 6 are respectively for implementing the corresponding flows in the method embodiments in fig. 1 to 5. Reference may be made specifically to the description of the above method embodiments, and a detailed description is omitted here where appropriate to avoid repetition.
As shown in fig. 7, an embodiment of the present application provides an electronic device 700, including: a processor 710, a memory 720 and a bus 730, wherein the processor is connected to the memory through the bus, and the memory stores computer readable instructions, which when executed by the processor, are used for implementing the method according to any of the embodiments described above, and in particular, refer to the description of the embodiments of the method, and the detailed description is omitted here to avoid redundancy.
Wherein the bus is used for realizing direct connection communication of the components. The processor in the embodiment of the present application may be an integrated circuit chip having signal processing capability. The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The Memory may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Read Only Memory (EPROM), an electrically Erasable Read Only Memory (EEPROM), and the like. The memory has stored therein computer readable instructions that, when executed by the processor, perform the methods described in the above embodiments.
It will be appreciated that the configuration shown in fig. 7 is merely illustrative and may include more or fewer components than shown in fig. 7 or have a different configuration than shown in fig. 7. The components shown in fig. 7 may be implemented in hardware, software, or a combination thereof.
Embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a server, the method in any of the above-mentioned all embodiments is implemented, which may specifically refer to the description in the above-mentioned method embodiments, and in order to avoid repetition, detailed description is appropriately omitted here.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made to the present application by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application. It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined or explained in subsequent figures.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A method of identifying anomalous regions of deep sea mineral production, the method comprising:
acquiring a pre-classified image and geochemical data corresponding to the pre-classified image, wherein the pre-classified image is obtained by preliminarily identifying an abnormal area, and the geochemical data is obtained by detecting deep sea mineral products in a target sea area;
and inputting the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification to obtain a target abnormal region classification result so as to finish further detection of the target sea area.
2. The method of claim 1, wherein the pre-classified image comprises a normal region and the abnormal region, the normal region having a content value of each chemical element in the geochemical data within a set threshold range, and the abnormal region having a content value of at least some chemical elements not within the set threshold range;
before the inputting the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification and obtaining a target abnormal region classification result, the method further comprises the following steps:
obtaining a pre-classified sample image;
inputting at least a pre-classified sample image and geochemical data corresponding to the pre-classified sample image into an abnormal area identification model to be trained for training to obtain the target abnormal area identification model;
and the abnormal area marked by the label in the pre-classified sample image belongs to a sub-area of the abnormal area in the pre-classified image.
3. The method of claim 2, wherein prior to said obtaining a pre-classified sample image, the method further comprises:
dividing the pre-classified image into M regions, wherein M is an integer greater than 1;
confirming that the M areas comprise N abnormal areas, wherein N is an integer greater than or equal to 1, and N is less than or equal to M;
and marking each abnormal region in the N abnormal regions as an enriched region or a depleted region according to the content of any chemical element included in each abnormal region, and obtaining the pre-classified sample image, wherein the content of the chemical element in the enriched region is greater than a first threshold value, the content of the chemical element in the depleted region is less than a second threshold value, and the second threshold value is less than the first threshold value.
4. The method of any of claims 1-3, wherein prior to said acquiring a pre-classified image and geochemical data corresponding to the pre-classified image, the method further comprises:
acquiring original mineral data, wherein the original mineral data is geochemical data corresponding to deep sea minerals of the target sea area;
dividing the original mineral data into a plurality of groups according to the attribute of each chemical element in the original mineral data to obtain a target classification result, wherein one group comprises at least one chemical element;
and according to the target classification result, preliminarily identifying a normal area and the abnormal area corresponding to the original mineral data to obtain the pre-classification image.
5. The method as claimed in claim 4, wherein the step of dividing the raw mineral data into a plurality of groups according to the attribute of each chemical element in the raw mineral data to obtain a target classification result comprises:
dividing the original mineral data into a plurality of groups through a multi-fractal algorithm to obtain a first classification result;
performing principal component analysis on the original mineral data, and dividing the original mineral data into a plurality of groups to obtain a second classification result;
performing cluster analysis on the original mineral data, dividing the original mineral data into a plurality of groups, and obtaining a third classification result;
and confirming that the first classification result, the second classification result and the third classification result are the same to obtain the target classification result.
6. The method of claim 5, wherein the dividing the original mineral data into a plurality of groups by a multi-fractal algorithm to obtain a first classification result comprises:
calculating a multi-fractal spectrum curve corresponding to each chemical element;
and dividing the original mineral data into a plurality of groups according to the fractal spectrum curve to obtain the first classification result.
7. A system for deep sea mineral anomaly identification, the system comprising:
the geochemical data acquisition equipment is used for detecting the deep-sea mineral products of the target sea area to obtain the original mineral product data of the target sea area;
a server configured to perform the method of any one of claims 1-6 on the basis of raw mineral data to obtain a target abnormal area classification result;
and the deep sea exploration equipment is used for further detecting the target sea area according to the target abnormal area division result.
8. An apparatus for identifying a deep sea mineral anomaly zone, the apparatus comprising:
the acquisition module is configured to acquire a pre-classified image and geochemical data corresponding to the pre-classified image, wherein the pre-classified image is obtained by preliminarily identifying an abnormal area, and the geochemical data is obtained by detecting deep sea mineral products of a target sea area;
and the identification module is configured to input the pre-classified image and the geochemical data into a target abnormal region identification model for re-identification, and obtain a target abnormal region division result so as to finish further detection of the target sea area.
9. An electronic device, comprising: a processor, memory, and a bus;
the processor is coupled to the memory via the bus, the memory storing computer readable instructions for implementing the method of any one of claims 1-6 when the computer readable instructions are executed by the processor.
10. A computer-readable storage medium, having stored thereon a computer program which, when executed, implements the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210605551.5A CN114776304B (en) | 2022-05-30 | 2022-05-30 | Method, device, equipment and medium for identifying abnormities of deep sea mineral products |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210605551.5A CN114776304B (en) | 2022-05-30 | 2022-05-30 | Method, device, equipment and medium for identifying abnormities of deep sea mineral products |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114776304A true CN114776304A (en) | 2022-07-22 |
| CN114776304B CN114776304B (en) | 2022-11-04 |
Family
ID=82420458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210605551.5A Active CN114776304B (en) | 2022-05-30 | 2022-05-30 | Method, device, equipment and medium for identifying abnormities of deep sea mineral products |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114776304B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116821651A (en) * | 2023-05-04 | 2023-09-29 | 广州海洋地质调查局 | Method and device for setting natural gas hydrate distant view zone and finely identifying ore body |
| CN117609848A (en) * | 2023-09-19 | 2024-02-27 | 昆明理工大学 | Mineral resource classification prediction method and system based on deep learning |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2003100678A4 (en) * | 2003-08-19 | 2003-11-20 | Opconsult Pty Ltd | Operations Resource Planning Software and Systems |
| CN112508041A (en) * | 2020-12-21 | 2021-03-16 | 苏州锐希信息技术有限公司 | Training method of neural network for spray control based on classification result label |
| CN112529112A (en) * | 2020-12-29 | 2021-03-19 | 中国地质科学院地质力学研究所 | Mineral identification method and device |
| CN112711646A (en) * | 2021-01-11 | 2021-04-27 | 中国地质大学（北京） | Geological information-based mining method and device, electronic equipment and storage medium |
| CN112785450A (en) * | 2020-12-30 | 2021-05-11 | 北京农业信息技术研究中心 | Soil environment quality partitioning method and system |
| CN112967766A (en) * | 2021-02-04 | 2021-06-15 | 辽宁工程技术大学 | Ore searching calibration method and system for element geochemical data screening |
| CN113361638A (en) * | 2021-07-01 | 2021-09-07 | 中国石油大学(北京) | Complex reservoir lithology identification method, device, equipment and storage medium |
| US20210279659A1 (en) * | 2020-03-09 | 2021-09-09 | Trimble Inc. | Construction site productivity capture using computer vision |
| WO2021197782A1 (en) * | 2020-03-31 | 2021-10-07 | Abb Schweiz Ag | Data processing for industrial machine learning |
| CN113488117A (en) * | 2019-04-27 | 2021-10-08 | 烟台市牟金矿业有限公司 | Deep gold deposit mining and prospecting method with deep learning ability |
| AU2021106209A4 (en) * | 2021-08-20 | 2021-10-28 | Changsha University Of Science & Technology | Photovoltaic power generation three-layer screening power theft identification method based on feature mining |
| CN113836128A (en) * | 2021-09-24 | 2021-12-24 | 北京拾味岛信息科技有限公司 | Abnormal data identification method, system, equipment and storage medium |
| CN113946950A (en) * | 2021-10-12 | 2022-01-18 | 青海省第三地质勘查院 | Method for rapidly delineating target area of gold prospecting |
-
2022
- 2022-05-30 CN CN202210605551.5A patent/CN114776304B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2003100678A4 (en) * | 2003-08-19 | 2003-11-20 | Opconsult Pty Ltd | Operations Resource Planning Software and Systems |
| CN113488117A (en) * | 2019-04-27 | 2021-10-08 | 烟台市牟金矿业有限公司 | Deep gold deposit mining and prospecting method with deep learning ability |
| US20210279659A1 (en) * | 2020-03-09 | 2021-09-09 | Trimble Inc. | Construction site productivity capture using computer vision |
| WO2021197782A1 (en) * | 2020-03-31 | 2021-10-07 | Abb Schweiz Ag | Data processing for industrial machine learning |
| CN112508041A (en) * | 2020-12-21 | 2021-03-16 | 苏州锐希信息技术有限公司 | Training method of neural network for spray control based on classification result label |
| CN112529112A (en) * | 2020-12-29 | 2021-03-19 | 中国地质科学院地质力学研究所 | Mineral identification method and device |
| CN112785450A (en) * | 2020-12-30 | 2021-05-11 | 北京农业信息技术研究中心 | Soil environment quality partitioning method and system |
| CN112711646A (en) * | 2021-01-11 | 2021-04-27 | 中国地质大学（北京） | Geological information-based mining method and device, electronic equipment and storage medium |
| CN112967766A (en) * | 2021-02-04 | 2021-06-15 | 辽宁工程技术大学 | Ore searching calibration method and system for element geochemical data screening |
| CN113361638A (en) * | 2021-07-01 | 2021-09-07 | 中国石油大学(北京) | Complex reservoir lithology identification method, device, equipment and storage medium |
| AU2021106209A4 (en) * | 2021-08-20 | 2021-10-28 | Changsha University Of Science & Technology | Photovoltaic power generation three-layer screening power theft identification method based on feature mining |
| CN113836128A (en) * | 2021-09-24 | 2021-12-24 | 北京拾味岛信息科技有限公司 | Abnormal data identification method, system, equipment and storage medium |
| CN113946950A (en) * | 2021-10-12 | 2022-01-18 | 青海省第三地质勘查院 | Method for rapidly delineating target area of gold prospecting |
Non-Patent Citations (3)
| Title |
|---|
| 张焱等: "多重分形与地质统计学方法在粤北刘家山地区矿化指示中的应用", 《地学前缘》 * |
| 张焱等: "多重分形法在琼东南盆地北礁凹陷砂体展布规律研究中的应用", 《海相油气地质》 * |
| 张焱等: "广东庞西垌地区地球化学组合异常识别与提取", 《地球学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116821651A (en) * | 2023-05-04 | 2023-09-29 | 广州海洋地质调查局 | Method and device for setting natural gas hydrate distant view zone and finely identifying ore body |
| CN117609848A (en) * | 2023-09-19 | 2024-02-27 | 昆明理工大学 | Mineral resource classification prediction method and system based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114776304B (en) | 2022-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114776304B (en) | Method, device, equipment and medium for identifying abnormities of deep sea mineral products | |
| Zhu et al. | Intelligent logging lithological interpretation with convolution neural networks | |
| US20230177347A1 (en) | Method and apparatus for detecting defect pattern on wafer based on unsupervised learning | |
| CN110210475B (en) | License plate character image segmentation method based on non-binarization and edge detection | |
| CN111738351B (en) | Model training method and device, storage medium and electronic equipment | |
| CN106295498B (en) | Remote sensing image target area detection device and method | |
| CN103473786B (en) | Gray level image segmentation method based on multi-objective fuzzy clustering | |
| CN108171119B (en) | SAR image change detection method based on residual error network | |
| CN105574550A (en) | Vehicle identification method and device | |
| CN104102929A (en) | Hyperspectral remote sensing data classification method based on deep learning | |
| CN109784383B (en) | Rail crack identification method based on graph domain feature and DS evidence theory fusion | |
| CN113344050B (en) | Lithology intelligent recognition method and system based on deep learning | |
| CN106557740B (en) | The recognition methods of oil depot target in a kind of remote sensing images | |
| CN105426441B (en) | A kind of automatic preprocess method of time series | |
| CN109118766A (en) | A kind of colleague's vehicle discriminating method and device based on traffic block port | |
| CN113516650B (en) | Circuit board hole plugging defect detection method and device based on deep learning | |
| WO2017101225A1 (en) | Trademark graph element identification method, apparatus and system, and computer storage medium | |
| CN115131580B (en) | Space target small sample identification method based on attention mechanism | |
| Shafiq et al. | The role of visual saliency in the automation of seismic interpretation | |
| Sasirekha et al. | Enhanced techniques for PDF image segmentation and text extraction | |
| Pessa et al. | Mapping images into ordinal networks | |
| CN117058446B (en) | Intelligent identification description method, system and storage medium for drilling core characteristics | |
| CN106951924B (en) | Seismic coherence body image fault automatic identification method and system based on AdaBoost algorithm | |
| CN106548195A (en) | A kind of object detection method based on modified model HOG ULBP feature operators | |
| CN104112145A (en) | Facial expression recognition method based on PWLD and D-S evidence theory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |